Currently, a company can have a network with numerous computers on it, all within multiple networks. The computer system that monitors the company's networks has sensors attached to some of their networks to detect the devices, operating systems, and the like on the networks. However, the sensors are not necessarily one per network. Furthermore, it is not necessarily apparent which one sensor is physically closest to which network.
In order to obtain a reliably accurate representation of the systems on a particular network and what the systems are doing (such as the servers and clients they are running and who they are connecting to), a user needs the information from the sensor which is closest to that network because the closest sensor has the best view of that network.
Currently, users manually specify which sensor is assigned to which network based on their own understanding of their network's topology and where their sensors are located.